The field of this invention relates to an amplifier circuit, an integrated circuit and a radio frequency communication unit comprising such an amplifier circuit and/or integrated circuit. The invention is particularly applicable to, but not limited to, a variable gain low noise amplifier circuit.
In the field of radio frequency (RF) communication receivers, the main task of the receiver front-end circuit is to process a signal that is received by an antenna coupled to the receiver front-end circuit in such a manner that it can be more easily processed by subsequent receiver circuits, for example, demodulation circuitry. Typically, such front-end circuits comprise low noise amplifier (LNA) circuitry for amplifying the received RF signal, and mixer circuitry arranged to perform frequency translation of the amplified radio frequency signal to a lower intermediate or baseband frequency. The intermediate/baseband frequency signal may then be filtered to remove interfering signals, etc.
Since the frequency of the intermediate or baseband signal output by the mixer circuitry is typically much lower than the carrier frequency (fRF) for the received RF signal, all stages within the receive chain subsequent to the mixer circuitry operates at low or baseband frequencies. Furthermore, due to the amplification provided by the LNA circuitry in front of the mixer circuitry, and by the mixer circuitry itself (if active mixers are used), the signal levels following the mixer circuitry are also larger than the signal level of the received RF signal. Accordingly, these low frequency/high signal level characteristics allow the use of a large variety of circuit techniques for the implementation of the stages within the receive chain following the front-end circuitry.
However, due to the high operating frequencies and the low signal levels of the received RF signal, only a very limited number of circuit techniques may be used to successfully implement the front-end circuitry that comprises the LNA circuitry and the mixer circuitry. The primary challenge in the design of an LNA circuit is to minimise noise. However, LNA circuitry within an RF receiver also has to provide a sufficiently large gain, a well defined input impedance, and has to introduce very little distortion (e.g. the performance of the LNA should be designed to be as linear as possible). Hence, one of the most important LNA linearity metrics is the IP3 (third order intercept point).
Referring now to FIG. 1, there is illustrated an example of a known LNA circuit topology 100, comprising an inductively degenerated amplifier. The LNA circuit topology 100 illustrated in FIG. 1 exploits the voltage gain provided by a series RLC resonance circuit to boost the voltage appearing between the gate and the source of the input device. This voltage amplification provides two advantages: firstly it provides amplification before the first noisy component of the amplifier, namely transistor M1 110; and secondly the effective transconductance of the amplifier input stage is increased by a factor ‘Q’ compared to the transconductance of transistor M1 110, where ‘Q’ is the quality factor of the input series resonance. The effect of providing such amplification before the first noisy component of the amplifier is a net reduction in the noise contributed by the amplifier over the total noise appearing at the output of the amplifier. Furthermore, the consequence of the increase of the effective transconductance of the amplifier input stage is a reduced current consumption for a given desired gain.
However, since the input stage is built around a resonant circuit, the input stage operates over relatively narrow bandwidths and, thus, has to be tuned differently for different frequency bands. In order to accommodate a large dynamic range, such as that required for modern communication receivers, the LNA circuitry is typically required to provide two or more gain settings. For the amplifier topology illustrated in FIG. 1, programmable gain settings are implemented by way of splitting the signal current using cascode transistors M2a 120 and M2b 130 such that, in all but the maximum gain setting, only part of the signal current reaches the output of the amplifier.
A problem with this approach is that it is inefficient in terms of current consumption, particularly at low gain settings. Accordingly, a desirable feature would be to be able to reduce the current consumption in the low gain settings. However, implementing any form of current reduction technique would change the transconductance of transistor M1 110. Since the input impedance of the amplifier topology 100 at resonance is real, and is proportional to the transconductance of transistor M1 110, such a current reduction would result in a change in the input impedance of the amplifier, which would cause a mismatch with, for example, an antenna coupled thereto.
A further problem with the amplifier topology 100 of FIG. 1 is that it exhibits a poor linearity performance. The voltage amplification provided by the input resonance circuit increases the gate-source voltage swing of transistor M1 110. Whilst this may be beneficial in terms of noise, it also increases the distortion introduced by transistor M1 110.
An alternative example of a known LNA circuit topology comprises a common-gate configuration. A problem with a traditional common-gate amplifier topology is that the theoretical best noise figure (NF) achievable is limited to 2.2 dB. The achievable noise figure is limited by the fact that the transconductance of the input device not only defines the noise characteristic of the amplifier, but it also determines its input impedance. A better noise figure can typically only be achieved by using reactive impedance transformations. This circuit configuration is therefore only used in receivers with relatively relaxed noise requirements. However, FIG. 2 illustrates an example of a recently proposed common-gate amplifier topology 200 in which the noise performance of the common-gate stage is improved. For the illustrated example, a common-source stage, comprising transistors Mc1b 210 and Mc2b 220, is connected in parallel with the common-gate stage, comprising transistors M1 230 and M2 240. If the transistors are properly sized, the noise of the common-gate transistor appears as a common-mode signal at the output of the amplifier, and can therefore be suppressed. The main noise contributor is then the common-source stage, which can be designed to have a higher transconductance than its common-gate counterpart. The higher transconductance common-source stage, together with the cancelling of the noise generated by the common-gate stage, result in an amplifier with an improved noise figure. However, the noise performance of such an amplifier topology 200 of FIG. 2 is still unable to match that of the inductively degenerated amplifier topology 100 of FIG. 1.
Nevertheless, an advantage of the amplifier topology 200 of FIG. 2 is that it converts a single ended input signal into a differential signal at the input of amplifier. A differential signal enables improved dynamic range, reduced sensitivity to supply voltage and substrate noise, improved isolation, etc. within, for example, a receiver chain of which the amplifier forms a part.
The input impedance of the amplifier topology 200 of FIG. 2 is broadband and is equal to the reciprocal of the transconductance of transistor M2 240. Accordingly, in the same manner as for the inductively degenerated amplifier of FIG. 1, the current cannot be reduced in the low gain modes, as this would modify the input impedance of the amplifier circuit. Gain control is therefore usually implemented with the help of cascode transistors in the same manner as described for the inductively degenerated amplifier of FIG. 1.
Referring now to FIG. 3, there is illustrated a further example of an amplifier topology 300 that is suitable for the implementation of an LNA, where the amplifier topology 300 comprises a shunt-shunt feedback amplifier. However, this configuration is not popular for the implementation of highly integrated receivers for mobile applications for two main reasons. Firstly, for proper operation the transconductance of transistor M1 310 has to be quite large (>100 mS), resulting in the amplifier, and in particular implementations comprising MOSFETs, being power hungry. Secondly, no straightforward way of implementing various gain settings has been proposed, as both the gain and the input impedance of the amplifier are functions of the feedback resistor RF 320, of the load resistor RL 330 and of the transconductance of M1 in a non-trivial way.
In addition to the above identified short comings of the prior art topologies, analogue circuits comprising components, such as inductors, are unable to scale and provide comparable improvements in integrated circuit manufacturing processes in the same manner as digital circuits. Instead, scaling of analogue circuits must be achieved by innovation and new design and circuit techniques.
Thus, a need exists for an improved amplifier circuit, integrated circuit and radio frequency communication unit that may alleviate one or more of the aforementioned problems of known amplifier circuits.